The present invention relates to a system for training cognitive capabilities within a cognitive solution, and more specifically, to providing for effective classification of contextually related data for the cognitive solution.
Ground truth is a term used in various fields to refer to information provided by direct observation. In machine learning, the term ground truth refers to the accuracy of the training data set's classification for supervised learning techniques. It involves a user that tells the machine learning algorithm what to lean. The user labels the training data examples to be classified by the machine learning algorithm. A trained machine learning model will be limited by the accuracy of the ground truth data used for training.